Study of composite interface strength and crack Bily, Mollie A.

Study of composite interface strength and crack Bily, Mollie A.
Calhoun: The NPS Institutional Archive
Theses and Dissertations
Thesis Collection
2009-09
Study of composite interface strength and crack
growth monitoring using carbon nanotubes
Bily, Mollie A.
Monterey, California. Naval Postgraduate School
http://hdl.handle.net/10945/4644
NAVAL
POSTGRADUATE
SCHOOL
MONTEREY, CALIFORNIA
THESIS
STUDY OF COMPOSITE INTERFACE STRENGTH
AND CRACK GROWTH MONITORING USING CARBON
NANOTUBES
by
Mollie A. Bily
September 2009
Thesis Advisor:
Second Reader:
Young W. Kwon
Randall D. Pollak
Approved for public release; distribution is unlimited
THIS PAGE INTENTIONALLY LEFT BLANK
REPORT DOCUMENTATION PAGE
Form Approved OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instruction,
searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send
comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to
Washington headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA
22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188) Washington DC 20503.
1. AGENCY USE ONLY (Leave blank)
2. REPORT DATE
3. REPORT TYPE AND DATES COVERED
September 2009
Master’s Thesis
4. TITLE AND SUBTITLE Study of Composite Interface Strength and Crack
5. FUNDING NUMBERS
Growth Monitoring Using Carbon Nanotubes
6. AUTHOR(S) Mollie A. Bily
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
8. PERFORMING ORGANIZATION
Naval Postgraduate School
REPORT NUMBER
Monterey, CA 93943-5000
9. SPONSORING /MONITORING AGENCY NAME(S) AND ADDRESS(ES)
10. SPONSORING/MONITORING
N/A
AGENCY REPORT NUMBER
11. SUPPLEMENTARY NOTES The views expressed in this thesis are those of the author and do not reflect the official policy
or position of the Department of Defense or the U.S. Government.
12a. DISTRIBUTION / AVAILABILITY STATEMENT
12b. DISTRIBUTION CODE
Approved for public release; distribution is unlimited
13. ABSTRACT (maximum 200 words)
Interface strength of woven fabric composite layers was studied using Mode II fracture strength testing. Both carbon
fiber and glass fiber composites were used with the vinyl ester resin. First, the single-step cured (i.e., co-cured)
composite interface strength was compared to that of the two-step cured interface as used in the scarf joint technique.
The test results showed that the two-step cured interface was as strong as the co-cured interface, and the former had
even higher fracture toughness than the latter. The second study applied carbon nanotubes to the composite interface
using the two-step curing technique. Mode II fracture testing was performed for the interface containing carbon
nanotubes. The results indicated great improvement of the interface fracture toughness due to carbon nanotubes.
Finally, a study was conducted to detect interface crack growth using the carbon nanotubes introduced at the interface.
Because carbon nanotubes have high electric conductivity, the electric resistance was measured though the interface.
As the interface crack grew under a loading, there was a gradual increase of electric resistance. As a result, the change
of electric resistance in terms of crack length change was determined. The study showed that using carbon nanotubes
at a critical composite interface would not only strengthen its fracture toughness but also detect crack growth.
14. SUBJECT TERMS Carbon Nanotubes, CNTs, Carbon Fiber Composite, Fiberglass Composite,
Crack Propagation, Mode II, Health Monitoring, Resistance Testing
17. SECURITY
CLASSIFICATION OF
REPORT
Unclassified
18. SECURITY
CLASSIFICATION OF THIS
PAGE
Unclassified
NSN 7540-01-280-5500
15. NUMBER OF
PAGES
93
16. PRICE CODE
19. SECURITY
20. LIMITATION OF
CLASSIFICATION OF
ABSTRACT
ABSTRACT
Unclassified
UU
Standard Form 298 (Rev. 2-89)
Prescribed by ANSI Std. 239-18
i
THIS PAGE INTENTIONALLY LEFT BLANK
ii
Approved for public release; distribution is unlimited
STUDY OF COMPOSITE INTERFACE STRENGTH AND CRACK GROWTH
MONITORING USING CARBON NANOTUBES
Mollie A. Bily
Lieutenant, United States Navy
B.S., United States Naval Academy, 2003
Submitted in partial fulfillment of the
requirements for the degree of
MASTER OF SCIENCE IN MECHANICAL ENGINEERING
from the
NAVAL POSTGRADUATE SCHOOL
September 2009
Author:
Mollie A. Bily
Approved by:
Young W. Kwon
Thesis Advisor
Randall D. Pollak
Second Reader
Knox T. Millsaps
Chairman, Department of Mechanical and Astronautical
Engineering
iii
THIS PAGE INTENTIONALLY LEFT BLANK
iv
ABSTRACT
Interface strength of woven fabric composite layers was studied using Mode II
fracture strength testing. Both carbon fiber and glass fiber composites were used with the
vinyl ester resin. First, the single-step cured (i.e., co-cured) composite interface strength
was compared to that of the two-step cured interface as used in the scarf joint technique.
The test results showed that the two-step cured interface was as strong as the co-cured
interface, and the former had even higher fracture toughness than the latter. The second
study applied carbon nanotubes to the composite interface using the two-step curing
technique. Mode II fracture testing was performed for the interface containing carbon
nanotubes. The results indicated great improvement of the interface fracture toughness
due to carbon nanotubes. Finally, a study was conducted to detect interface crack growth
using the carbon nanotubes introduced at the interface. Because carbon nanotubes have
high electric conductivity, the electric resistance was measured though the interface. As
the interface crack grew under a loading, there was a gradual increase of electric
resistance. As a result, the change of electric resistance in terms of crack length change
was determined. The study showed that using carbon nanotubes at a critical composite
interface would not only strengthen its fracture toughness but also detect crack growth.
v
THIS PAGE INTENTIONALLY LEFT BLANK
vi
TABLE OF CONTENTS
I.
INTRODUCTION........................................................................................................1
A.
BACKGROUND ..............................................................................................1
B.
LITERATURE REVIEW ...............................................................................3
C.
OBJECTIVES ..................................................................................................4
II.
COMPOSITE SAMPLE CONSTRUCTION............................................................7
A.
SAMPLE SPECIFICATION ..........................................................................7
B.
MATERIALS ...................................................................................................8
C.
VACUUM-ASSISTED RESI
NT
RANSFER MOLDING
TECHNIQUE ...................................................................................................8
III.
PHASES OF RESEARCH ........................................................................................21
A.
PHASE I: FAMILIARIZATION .................................................................21
B.
PHASE II: CO-CURED VS. TWO-STEP CURED....................................21
C.
PHASE III: CARBON COMPOSITE RESISTANCE TESTING ............21
D.
PHASE IV: FIBERGLASS COMPOSITE RESISTANCE TESTING ....22
E.
PHASE V: RESIST ANCE RELIABILITY AND CRACK GROWTH
RELATIONSHIP TESTING ........................................................................22
IV.
TESTING....................................................................................................................25
A.
EQUIPMENT.................................................................................................25
B.
PROCEDURE ................................................................................................26
C.
CALCULATIONS .........................................................................................28
V.
RESULTS AND DISCUSSION ................................................................................29
A.
PHASE I: FAMILIRIZATION ....................................................................29
B.
PHASE II: CO-CURED VS. TWO-STEP CURED....................................29
C.
PHASE III: CARBON COMPOSITE RESITANCE TESTING ..............34
D.
PHASE IV: FIBERGLASS COMPOSITE RESISTANCE TESTING ....41
E.
PHASE V: RESIST ANCE RELIABILITY AND CRACK GROWTH
RELATIONSHIP TESTING ........................................................................49
VI.
CONCLUSIONS AND RECOMMENDATIONS...................................................57
APPENDIX A: T WO-STEP CURED AND CO-CURED CRITICAL STRAI N
ENERGY RELEASE RATES (GII)..........................................................................59
Two-Step Cured .............................................................................................59
Co-Cured ........................................................................................................59
APPENDIX B: CARBON COMPOSIT
E WITH CNT RESI STANCE DAT A
PHASE III...................................................................................................................61
APPENDIX C: PURE CARBON COMPOSITE RESISTANCE DATA PHASE III ....63
APPENDIX D: CARBON COMPOS
ITE WI TH AND WITHOUT CNT
CRITICAL STRAIN ENERGY RELEASE RATES (GII) ....................................65
With CNT .......................................................................................................65
vii
Without CNT..................................................................................................65
APPENDIX E: FIB ERGLASS COMPOS ITE WITH CNT RE SISTANCE DAT A
PHASE IV...................................................................................................................67
APPENDIX F: FI BERGLASS COMPOSITE WIT H AND WITHOUT CNT
CRITICAL STRAIN ENERGY RELEASE RATES (GII) ....................................69
With CNT .......................................................................................................69
Without CNT..................................................................................................69
APPENDIX G: FIBERGLAS S COMPOS ITE WITH CNT RESI STANCE DAT A
PHASE V ....................................................................................................................71
APPENDIX H: CARBON COMPOSIT
E WITH CNT RESI STANCE DAT A
PHASE V ....................................................................................................................73
LIST OF REFERENCES ......................................................................................................77
INITIAL DISTRIBUTION LIST .........................................................................................79
viii
LIST OF FIGURES
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.
Figure 10.
Figure 11.
Figure 12.
Figure 13.
Figure 14.
Figure 15.
Figure 16.
Figure 17.
Figure 18.
Figure 19.
Figure 20.
Figure 21.
Figure 22.
Figure 23.
Figure 24.
Figure 25.
Figure 26.
Figure 27.
Figure 28.
Figure 29.
Figure 30.
Figure 31.
Figure 32.
Figure 33.
Figure 34.
Figure 35.
Figure 36.
Figure 37.
Figure 38.
Figure 39.
Figure 40.
Figure 41.
Figure 42.
Single-Walled Carbon Nanotube .......................................................................2
Multi-Walled Carbon Nanotube ........................................................................2
Sample Geometry...............................................................................................7
Bottom Layer of Distribution Media used for Co-Cured Samples ....................9
Peel Ply Laid on Top of Distribution Media for Co-Cured Samples...............10
Bottom Five Layers of a Sample .....................................................................11
Peel Ply and Distribution Media on Top of Stacked Fiber Layers ..................11
Gage Board and Resin Trap.............................................................................12
Spiral Tubing Used at the Top and Bottom of Sample Set-up .......................12
Vacuum Tape Used to Seal the Sample Setup.................................................13
Rolling Out the Plastic Sheet Used to Form the Vacuum Bag ........................14
Sample Setup under Vacuum...........................................................................14
Resin at Inlet with Bubbles after Mixing.........................................................15
Resin Running through a Sample Evenly ........................................................16
Resin Completely through a Sample................................................................16
Bottom Layer of Double-Cure Sample Covered With CNTs..........................17
Teflon Layer Used to Build Initial Crack in Sample .......................................18
Remaining Fiber Material Stacked on Top of Bottom Plate............................19
INSTRON Mode II Test Setup ........................................................................25
Fluke 8840A Multi-Meter and INSTRON Mode II Test Setup.......................26
Diagram of Three-point Bending Test for Mode II .........................................27
Picture of Three-point Bending Test for Mode II............................................27
Normalized Average Values of GII for Phase II ..............................................31
Crack Propagation Path for a Co-Cured Coupon.............................................32
Crack Propagation Path for a Two-Step Cured Coupon..................................32
Surface Crack Propagation Path for a Co-Cured Coupon ...............................33
Surface Crack Propagation Path for a Two-Step Cured Coupon.....................33
Carbon Fiber Mode II Resistance Testing Bent Position.................................35
Mode II Graph of Carbon Composites With CNT...........................................38
Mode II Graph of Carbon Composites Without CNT .....................................38
Normalized Average Values of GII for Phase III .............................................39
Surface Crack Propagation Path of Carbon Composite Without CNT............40
Surface Crack Propagation Path of Carbon Composite With CNT.................40
Fiberglass Coupon With Gaps in the Layer of CNTs ......................................42
Fiberglass Coupon With Continuous Layer of CNTs......................................42
Normalized Average Values of GII for Phase IV.............................................44
Mode II Graph of Fiberglass Composites With CNT......................................45
Mode II Graph of Fiberglass Composites Without CNT.................................45
Fiberglass Composites Without CNT Path of Crack Propagation Drawing....46
Fiberglass Composites Without CNT Path of Crack Propagation Picture ......46
Fiberglass Composites With CNT Path of Crack Propagation........................47
Fiberglass Composites With CNT Path of Crack Propagation Picture............47
ix
Figure 43.
Figure 44.
Figure 45.
Figure 46.
Figure 47.
Figure 48.
Figure 49.
Figure 50.
Figure 51.
Figure 52.
Figure 53.
Figure 54.
Surface Crack Propagation Path of Fiberglass Composite With CNT ............48
Surface Crack Propagation Path of Fiberglass Composite Without CNT .......48
Carbon Composite Resistance vs. Crack Length Graph For All CNT
Coupons ...........................................................................................................51
Carbon Composite Coupon 1 Resistance vs. Crack Length Graph .................52
Carbon Composite Coupon 2 Resistance vs. Crack Length Graph .................52
Carbon Composite Coupon 3 Resistance vs. Crack Length Graph .................53
Carbon Composite Coupon 4 Resistance vs. Crack Length Graph .................53
Carbon Composite Coupon 5 Resistance vs. Crack Length Graph .................54
Carbon Composite Coupon 6 Resistance vs. Crack Length Graph .................54
Carbon Composite Coupon 8 Resistance vs. Crack Length Graph .................55
Carbon Composite Coupon 9 Resistance vs. Crack Length Graph .................55
Carbon Composite Coupon 10 Resistance vs. Crack Length Graph ...............56
x
ACKNOWLEDGMENTS
I would like to thank Dr. Young Kwon for his mentorship and patience during the
course of this research and throughout my graduate studies.
Major Randall Pollak, USAF, is also appreciated for providing guidance
particularly in the area of Carbon Nanotube Non-Destructive Testing.
Many thanks to Tom Christen and Chanman Park for their technical advice and
for giving up their valuable time to help choose, set up, and take down test equipment.
Thanks to the Air Force Office of Scientific Research and the Naval Surface
Warfare Center Carderock Division (NSWCCD) team for “Advanced Hull Materials &
Structures Technology (AHM&ST)” who provided crucial funding and materials.
Thank you to Integrated Composites for use of their equipment during the course
of my research.
Last, but certainly not least, I would like to thank my husband, John, for his
support and patience throughout my entire time here at the Naval Postgraduate School.
xi
THIS PAGE INTENTIONALLY LEFT BLANK
xii
I. INTRODUCTION
A. BACKGROUND
In the past decade, composite structures have been in the forefront of structural
research. In particular, the Department of Defense has looked at both carbon fiber and
fiberglass composites for use in construction of ship superstructures, submarine sails, and
structures of unmanned aircraft.1 Many promising steps have been taken to ensure that
composites are fully integrated into structural use. A recurring hindrance to successful
integration of composite structures is that of critical areas of stress concentration, and
their ability to withstand failure.
Carbon-Carbon bonds are one of the strongest of chemical bonds found in nature,
and are the basis for the strength of carbon nanotubes. Carbon nanotubes (CNTs) are
made of sp2 hybridized carbon bonds, with each atom joined to three neighbors creating a
hexagonal lattice structure like graphite.2 The lattice structure forms a tube with a nanosized diameter and can be several millimeters in length, as shown in Figure 1. The three
main classifications of CNTs used in modern research are single-walled, double-walled,
and multi-walled, meaning an inner cylinder lies within the outer cylinder as shown in
Figure 2.3 Although many strides have been made in the manufacturing of CNTs, they
are still quite expensive. CNTs have an extremely high elastic modulus (greater than 1
TPa), high tensile strengths (up to 63 GPa), and are extremely lightweight, making them
ideal for reinforcement of composite materials.4
1 A.P. Mouritz, E. Gellert, P. Burchill, and K. Challis, “Review of Advanced Composite Structures for
Naval Ships and Submarines,” Composite Structures 53 (2001): 21–41.
2R. Saito and M. S. Dresselhaus, Physical Properties of Carbon Nanotubes (Imperial College Press,
1998), 11–12.
3William D. Callister, Jr., Materials Science and Engineering: An Introduction (New York: John
Wiley and Sons, Inc, 2007), 433.
4P.J.F. Harris. “Carbon Nanotube Composites,” International Materials Review 49 (2004): 31.
1
Figure 1.
Single-Walled Carbon Nanotube5
Figure 2.
Multi-Walled Carbon Nanotube6
It has already been demonstrated that inclusion of CNTs in areas of high stress
concentration can increase a material’s ability to withstand stress at these critical areas.7
However, a secondary benefit could be the use of CNTs to monitor composite materials
5 The Venton Research Group. Development of Carbon Nanotube Modified Microelectrodes. n.d.
http://www.faculty.virginia.edu/ventongroup/nanotube.html (accessed September 9, 2009)
6 Live Journal. Definition of a Nanotube, March 12, 2009. http://fullerenes.livejournal.com/ (accessed
September 9, 2009).
7
Susan Faulkner, Study of Composite Joint Strength with Carbon Nanotube Reinforcement, Naval
Postgraduate School, MS thesis, September 2008, 1–42.
2
to detect damage at interfaces. Compared to metals, the failure of composites is much
more difficult to predict due to the accumulation of damage ultimately leading to failure.8
Since failure is often difficult to predict, employing a network of CNTs at a critical
juncture would provide a dual purpose for their inclusion in the composite material.
Composite materials would be strengthened, while simultaneously detecting interfacial
damage.
B. LITERAT
URE REVIEW
Many different studies have been conducted to determine the feasibility of
damage detection in composite materials through the use of CNTs. During one study a
CNT polymer material was used to manufacture a piezoresistive strain sensor for
structural health monitoring.
This sensor proved to have a linear symmetric strain
response under static and dynamic loading, however the CNTs were only included within
the sensor itself.9 One similar study conducted showed that multidirectional strains could
be measured using an isotropic film of CNTs placed on a four point probe. This probe
then could be moved around to different locations sensing a linear strain response in all
locations.10
Another study, however, focused solely on the use of CNTs as a replacement for
strain gauges.
This study placed semi-conductive multiwall CNT-fiberglass–epoxy
polymer composites under both tensile and cyclic loading to detect failure. It was shown
that the multiwall CNTs were able to outperform regular strain gauges in sensing
different types of failures. This was due to their ability to be interspersed within the
composite and, as a result, be more sensitive to the changing stress fields around them.11
8 I. Weber, and P. Schwartz, “Monitoring Bending Fatigue In Carbon-Fibre/Epoxy Composite
Strands: A Comparison Between Mechanical and Resistance Techniques,” Composites Science and
Technology 61 (2001): 849–853.
9 I. Kang, M.J. Schulz, J.H. Kim, V. Shanov, and D. Shi, “A Carbon Nanotube Strain Sensor for
Structural Health Monitoring,” Smart Materials and Structures, 15 (2006): 737–748.
10P. Dharap, Z. Li, S. Nagarajaiah, and Barrera, E.V, “Nanotube Film Based on Single-Wall Carbon
Nanotubes for Strain Sensing,” Nanotechnology, 15 (2004): 379–382.
11 M. Nofar, S.V. Hoa, and M.D. Pugh, “Failure Detection and Monitoring in Polymer Matrix
Composites Subjected to Static and Dynamic Loads Using Carbon Nanotube Networks,” Composites
Science and Technology (2009): 1–22.
3
Much work has been done to replace strain gauges, however limited research has
been conducted based on crack propagation and local application of CNTs. In one study
CNTs were first dispersed into a polymer matrix and then infiltrated into layers and
bundles of conventional fibers. This created a percolating network which was then used
as a sensor to detect the onset, nature and evolution of damage in advanced-polymerbased composites.12 A similar study demonstrated that a network of CNTs throughout the
composite material is an effective way to monitor fatigue-induced damage, as well as
opportunities for damage repair.13 Yet, another study showed that if a high aspect ratio
could be maintained throughout the entire network of CNTs, they could be highly
conductive within the structure allowing for damage detection.14
Each of these studies however, used a network of CNTs dispersed throughout the
composite base material to achieve damage detection.
These methods, although
successful in the detection of damage, still may not address the interfacial damage
mechanisms.
In order to achieve this type of damage detection, a layer of CNTs
percolated along the matrix surface is to be studied.
C. OBJECTIVES
The objective of this research is to advance the uses of CNTs within composite
materials. Two main areas of research will be looked at and studied closely to further the
implementation of CNTs as a local reinforcement.
Previous research showed that CNTs can increase the fracture toughness of the
composite interface significantly; however, only one assembly mode, two-step cured, was
used.15 The first objective of this research is to determine the critical strain energy
12 E.T. Thostenson and T. W. Chou, “Carbon Nanotube Networks: Sensing of Distributed Strain and
Damage for Life Prediction and Self Healing,” Advanced Materials, 18 (2006): 2837–2841.
13 W. Zhang, V. Sakalkar, and N. Koratkar, “In Situ Health Monitoring and Repair In Composites
Using Carbon Nanotube Additives,” Applied Physiscs Letters, 91(2007).
14Tsu-Wei Chou and Erik T. Thosetenson. “Carbon Nanotube/Vinyl Ester Nanocomposites for in Situ
Sensing,” September 17-29, 2008. University of Maryland University College, Adelphia, MD. Office of
Naval Research Solid Mechanics Program Review Meeting: Marine Composites and Sandwich Structures:
42–49.
15Faulkner, “Study of Composite Joint Strength with Carbon Nanotube Reinforcement,” 15–42.
4
release rate, G, and crack propagation characteristics of carbon fiber vinyl ester resin
composite during Mode II fractures for both single-step-cured (co-cured) and two-step
cured composite sample sets. This ultimately will help determine the optimum way to
manufacture composite materials with, and without, CNTs. If the data is the same, the
methods can be interchanged and allow for more flexibility in the composite material
assembly process.
Ideally, if CNTs are close enough together, as result of their conductive nature,
they can conduct electrical current. The second objective of this research is to exploit
this characteristic in order to determine if failure in a composite interface has occurred.
Failure of a composite would occur if the material has deformed enough that the CNTs
are no longer touching each other. In real-world applications, a procedure using current
to test for failure would be beneficial to operational units, and would provide a method
for real time monitoring, such as in situ health monitoring.
5
THIS PAGE INTENTIONALLY LEFT BLANK
6
II. COMPOSITE
A. SAMPLE
SAMPLE CONSTRUCTION
SPECIFICATION
Three different sample sets were constructed during this research. The first set
consisted of two types of resin-only carbon composite coupons: one set of coupons being
co-cured, and the other set being two-step cured. The second and third sample groups
also consisted of two coupon types per sample set.
One coupon group was fiber
composite with resin-only, while the other fiber composite group was CNT-reinforced.
The differences between the second and third sample sets were the type of base
composite material used, being carbon and fiberglass, respectively.
Each sample set consisted of the same basic coupon construction, with varying
parameters, and materials. All coupons had pre-existing cracks built into them in order to
represent an area of high-stress concentration. The first sample set of coupons is exactly
as depicted in Figure 3, whereas the second and third sample sets had stainless steel metal
sheets built into each end to allow for current to run through the sample sets. For these
two sample sets of coupons, the length of the crack was made sufficiently longer so that
the extra width of the thin piece of metal did not affect the test results.
Figure 3.
Where:
Sample Geometry16.
2L = length
h = thickness
a = initial crack length
16 Faulkner, “Study of Composite Joint Strength with Carbon Nanotube Reinforcement,” 19.
7
B. MATERI
ALS
Sample sets one and two were both constructed of TORAY T700CF carbon fiber
weave with a vinyl-ester matrix whose base was DERAKANE 510–A. The third sample
set also used DERKANE 510–A to create the base, but this time was made with
bidirectional fiberglass woven roving. Typically, fiberglass woven roving is categorized
by weight in ounces per square yard; for this research, 24-oz per square-yard woven
roving was used. Both the carbon fiber weave and the fiberglass woven were chosen
based on their current use in DoD structural projects.
In order to make the vinyl ester matrixes, the DERAKANE 510–A had to be
cured and hardened. The hardening chemicals used for this process are Methyl Ethyl
Ketone Peroxide (MEKP) and Cobalt Naphthenate (CoNap). MEKP was used to initiate
the chemical reaction to cure the DERAKAN 510–A, while CoNap was used to ensure
that the reaction occurred in the desired cure time. For this research, the desired cure
time was 60 minutes, which provided enough time for the DERAKANE to completely
penetrate all layers of the woven materials.
The above two hardeners work well if the ambient temperature is between 70˚F
and 80˚F, in which case the combination of hardeners was 1.25 weight percent MEKP
and 0.20 weight percent CoNap. For most of the research, the ambient temperature was
well below 70˚F and a third chemical, N-dimthylaniline (DMA), was needed to ensure a
cure time of 60 minutes. When DMA was used in combination with the previously stated
weight percentages for CoNap and MEKP, a total of 0.05 weight percent of DMA was
required. If DMA was not included at these low temperatures, cure times were much
longer than the desired 60 minutes.
C.
VACUUM-ASSISTED RESIN TRANSFER MOLDING TECHNIQUE
One technique for making composite materials in industry is Vacuum-Assisted
Resin Transfer Molding (VARTM), which was used in this thesis to construct the three
different sample sets required for testing. The VARTM process uses a vacuum to pull
resin through the many layers of fiber to ensure a uniform distribution of resin throughout
8
the sample. This technique was extremely beneficial when working with CNTs, as they
did not shift or move when the resin was run through the sample.
To begin making the two-step cured samples, a layer of peel ply was placed on a
piece of glass to allow for easy removal of the sample upon completion of the VARTM
process. The glass used must be at least 1.27 cm (0.5 in) thick, in order to be able to
withstand the extreme heat generated during the resin curing process. When making a
co-cured sample, a layer of distribution media is laid down first, covered by a layer of
peel ply, as shown in Figures 4 and 5.
Figure 4.
Bottom Layer of Distribution Media used for Co-Cured Samples
9
Figure 5.
Peel Ply Laid on Top of Distribution Media for Co-Cured Samples
Next, the sample size was chosen and the fiber materials were cut to the
appropriate size. For all samples, 10 layers of fabric were cut, five for the bottom layer,
and five for the top layer. The bottom five layers were then placed on top of the peel ply,
as shown in Figure 6. For the co-cure process, a Teflon film of thickness 0.0051 cm
(0.002 in) was placed partially on top of the bottom five layers in order to build a crack
into the sample. The last five layers of fiber material were evenly stacked on top of the
fiber material and Teflon already in place. Then another layer of peel ply, followed by a
piece of distribution media, was stacked on top of the complete co-cure sample. For the
double-cure sample, the bottom five layers were covered with the peel ply and
distribution media, as shown in Figure 7.
10
Figure 6.
Figure 7.
Bottom Five Layers of a Sample
Peel Ply and Distribution Media on Top of Stacked Fiber Layers
In order for the resin to be pulled through the fiber material, a Rietschel Thomas
Vacuum Pump model 2688CE44 was used. Tubing was hooked up to this pump and run
through a gauge board to a resin trap, as shown in Figure 8. The resin trap was used to
protect both the pump and gauge board from excess resin. From the resin trap, solid ½inch diameter plastic tubing was measured and cut to be used inside the vacuum bag as
the outlet for the resin. This same tubing was used to suck resin from the bottom of the
11
sample to the top. Attached to both the inlet and outlet tubes, and spread across the top
and bottom of the sample, was spiral tubing, as shown in Figure 9. This tubing allowed
for an even distribution of the resin throughout the sample.
Figure 8.
Figure 9.
Gage Board and Resin Trap
Spiral Tubing Used at the Top and Bottom of Sample Set-up
Once the tubing was assembled and secured, strips of vacuum bag tape were laid
out in a box shape around the sample stack. The strips were placed about 2 to 3 inches
12
from the sample stack, so as not to interfere with the resin being run through the sample.
The tape was used to hold the plastic sheet in place, which ultimately acted as a vacuum
bag, Figures 10 and 11. The plastic sheet was cut to fit the square box already made, and
was carefully rolled out onto the tape, Figures 11 and 12. The vacuum was turned on,
and the newly-created bag was thoroughly checked to make sure there were no leaks. If
there were to have been a leak in the bag, air bubbles would have entered both the bag
and the sample, making the sample unusable. Once it had been verified there were no
leaks, the vacuum was left on to ensure a continuous vacuum pressure throughout the rest
of the VARTM process.
Figure 10.
Vacuum Tape Used to Seal the Sample Setup
13
Figure 11.
Rolling Out the Plastic Sheet Used to Form the Vacuum Bag
Figure 12.
Sample Setup under Vacuum
While the vacuum was still running, the temperature was noted and the
appropriate amounts of resin and hardeners were mixed to ensure a 60-minute cure time.
Once mixed, the resin was transferred to the inlet of the vacuum bag and the inlet tube
was clamped to prevent the resin from flowing through the sample. As a result of mixing
and transferring the resin to a new bucket, small bubbles are formed throughout the resin,
Figure 13. Enough time, about ten to fifteen minutes, was allowed for these bubbles to
14
dissipate before running the resin through the sample. Again, these small bubbles, if
allowed to run through the sample, would have gotten caught and ruined the sample.
Figure 13.
Resin at Inlet with Bubbles after Mixing
After sufficient time had passed and no small bubbles could be seen in the resin,
the inlet tube was unclamped slowly to allow the resin to enter the vacuum bag. The
resin flowed evenly through the sample at a steady pace, as shown in Figure 14. The
resin was allowed to run all the way through the spiral tubing on the top, in order to
ensure all fibers were coated with the resin as in Figure 15. One aid used to ensure that
all fibers were covered with resin was the placement of the distribution media at the
beginning of the VARTM setup. When both a top and bottom layers were used the
bottom distribution media hung out the bottom of the sample by about ½ inch. The top
distribution media was then place under the top spiral tubing and even with the bottom of
the sample. This placement aided in sucking the resin up from the bottom of the sample,
through the middle, and out the top.
15
Figure 14.
Figure 15.
Resin Running through a Sample Evenly
Resin Completely through a Sample
As the resin started to cure, it became extremely hot and started to gel. When this
occurred, and all the layers were covered with resin, the resin inlet tubing was again
clamped to ensure no air was pulled into the sample. The time it took for this to happen
depended on the thickness and size of the sample, as well as the amount of resin and
hardeners used. The sample was left with the vacuum pump running until the sample
cured. If the resin and hardeners were mixed and added correctly, this was about 60
16
minutes. After this time, the pump was shut off, but the sample was left at least 12 hours
to ensure complete curing of the sample.
At this point, the co-cured sample was
complete and was taken to a water jet to get cut into the correct coupon size. For the twostep cured process more work was needed to complete the sample.
Since the bottom layer of the two-step cured sample was the only thing made the
first time through, the initial crack and top layer were then manufactured. To do this, the
first start step was to take the newly-made bottom layer, and sand the top surface with
100 grit sand paper in order to roughen the surface. Next, the sanded surface was cleaned
with acetone, in order to make sure that all sanded particles are removed. The acetone
was allowed to fully dry before continuing the VARTM process. When working with
CNTs, they were dispersed over the top of the entire sanded composite plate using
acetone, and again ensuring enough time was allowed for the acetone to dry, as shown in
Figure 16.
Figure 16.
Bottom Layer of Double-Cure Sample Covered With CNTs
For sample sets two and three, thin pieces of stainless steel plates were fastened to
the top and bottom of the sample, as shown above in Figure 16. The stainless steel was
needed to allow for a place to secure conductive test equipment to the sample and not
interfere with any other testing. For all other samples, this step was skipped.
17
Finally, the same steps as before were followed. Peel ply was laid on the glass
followed by the bottom composite plate. The crack was formed using the same Teflon
material as before and is shown in Figure 17. The previously-cut five pieces of fiber
material used to make the top plate were carefully stacked on top, Figure 18. More peel
ply was used, again followed by a piece of distribution media on top. Tubing was cut,
tape was laid out, and the vacuum bag was sealed and tested. The resin was then mixed,
allowed to sit while bubbles were popped, and then the resin was run through the sample.
The resin got hot, gelled, and 60 minutes later it was completely cured and the pump was
shut off. Again, the sample was given about 12 hours to sit and fully set. The two-step
cured sample was complete and was taken to be cut using a water jet.
Figure 17.
Teflon Layer Used to Build Initial Crack in Sample
18
Figure 18.
Remaining Fiber Material Stacked on Top of Bottom Plate
19
THIS PAGE INTENTIONALLY LEFT BLANK
20
III.
A.
PHASES OF RESEARCH
PHASE I: FAMILIARIZATION
Phase I consisted of a familiarization stage, during which samples were
constructed to simply learn the finer ins and outs of the VARTM technique. Several
samples were constructed, but only the last few were usable. The samples that did not
turn out were cut open and examined to help correct the problems.
The good samples
were cut into coupons and tested in order to learn how to use the test equipment, but no
data was collected.
B.
PHASE II: CO-CURED VS. TWO-STEP CURED
Phase II was conducted in order to determine the validity of using a co-cure
method versus a two-step cure method when making samples. This phase consisted of
two different sets of carbon composite samples that did not include CNTs. Samples were
cut into coupons 2.4 cm wide, 0.42 cm thick, and 17 cm long, based on applicable ASTM
Standards. The coupons were tested in Mode II and critical strain energy release rate, G,
was calculated.
C.
PHASE III: CARBON COMPOSITE RESISTANCE TESTING
Once Phase II was complete, two new carbon composite sample sets were
constructed. One set of samples was the same as the Phase II two-step cured samples,
while the other sample set included a layer of CNTs dispersed through the center of the
sample. CNTs surface concentration was 7.5 g/m2 and was dispersed using acetone. The
selection of CNTs surface concentration, as well as the selection of acetone as the
dispersing agent, was based on results from compression testing of CNTs reinforced scarf
joints conducted during previous research.17 Additionally, built into each sample set at
the far ends were thin pieces of stainless steel metal. This was used in order to prevent a
17Y. W. Kwon, R. Slaff, S. Bartlett, and T. Greene, “Enhancement of Composite Scarf Joint Interface
Strength through Carbon Nanotube Reinforcement,” Journal of Materials Science (2008): 1–9.
21
larger crack in the sample than was used in Phase II, but to allow for electrical testing to
be conducted. These sample sets were cut into the same size coupons as used in Phase II.
The purpose of this phase of research was to determine if a layer of CNTs could
be used to detect crack propagation making use of the CNTs’ electrical conductive
nature. Both sets of samples were tested in Mode II, while an electrical current was run
through them and the resistance was monitored. The resistance changes before, during,
and after Mode II testing were noted and critical strain energy release rate, G, was
calculated.
D.
PHASE IV: FIBERGLASS COMPOSITE RESISTANCE TESTING
Upon completion of Phase III, the exact same size sample sets used in Phase III
were constructed and cut into the same coupon sizes, except fiberglass was used as the
base composite material. To ensure the CNTs were allowed to touch one another, a
CNTs dispersion concentration of 10 g/m2 was used. The purpose of the phase was to
determine if the CNTs’ electrical conductivity could be exploited in even lesser
conductive composite materials. Ideally, even in low conductive materials, some current
will flow through the CNTs middle layer, allowing for crack propagation to be detected.
Both sets of samples were tested in Mode II, while an electrical current was run through
them and the resistance was monitored. The resistance changes before, during, and after
Mode II testing were noted and critical strain energy release rate, G, was calculated.
E.
PHASE V: RESIS TANCE RE LIABILITY AND CRACK GROWTH
RELATIONSHIP TESTING
This phase put both the carbon and fiberglass composites reinforced with CNT in
Phases III and IV through more tests. These tests were designed to determine the
reliability of the resistance readings collected in Phases III and IV, as well determine if
there is a relationship between the changes in crack length to the changes in resistance.
The first test used the previously cracked sample sets with CNTs from both
Phases III and IV, and slowly loaded them to a desired load prior to the point of further
22
crack propagation. The resistance readings were then read while under load, and then
upon unloading of the sample. This step was then repeated several times to determine the
consistency of the resistance readings.
For the second test, the crack length acquired during previous phases of research,
for each coupon, was measured along with the corresponding resistance reading. The
cracked coupon was then placed under a high enough load for the crack to propagate.
Upon propagation of the crack, and while still under load, a resistance reading was taken.
The load was removed and another resistance reading was measured. This procedure was
repeated until it was no longer possible to propagate the crack further. The resulting data
was used to determine relationships between change in crack length and change in
resistance readings.
23
THIS PAGE INTENTIONALLY LEFT BLANK
24
IV. TES TING
A. EQUIPME
NT
All tests were conducted using an Instron Tension/Compression Machine (Model
Number: 4507/4500), shown in Figure 19. All testing phases were conducted using a 10
kN load cell. Collection of data generated by the Instron Machine was done by a Series
IX computer software which was also used to control the Instron to achieve desired test
requirements. Additionally, for Phase III, IV and V the coupons were hooked up to a
Fluke 8840A Multi-meter as displayed in Figure 20. This device was used to measure the
resistance within the coupons throughout the entire test period. Data produced from this
machine was collected by hand at 30-second intervals.
Figure 19.
INSTRON Mode II Test Setup
25
Figure 20.
Fluke 8840A Multi-Meter and INSTRON Mode II Test Setup
B. PROCEDURE
In order to model a Mode II fracture, in which only shear force affects crack
propagation, each sample set was tested using a three-point bending test. This test was
chosen based on previous research conducted.
The setup used is shown in Figures 21 and 22. For all tests, the Instron held the
center support stationary, attached to the load cell, while the base supports were
incrementally moved up into the stationary support. The higher the base moved the
greater the load felt on the coupon became, resulting in higher shear stresses felt at the
crack tip. A plot of force versus displacement was provided from the Series IX computer
software and used to help calculate the Mode II critical strain energy release rate, GII.
26
Figure 21.
Figure 22.
Diagram of Three-point Bending Test for Mode II18
Picture of Three-point Bending Test for Mode II
Additionally, during Phase III, IV and V testing, the resistance of each coupon
was monitored. At the point of crack propagation, the resistance through the coupon was
annotated and compared to that of the initial resistance reading. The resistance was again
taken after the test had stopped and the coupon was still bent. Another reading was taken
after the coupon was removed from the Instron and returned to a load free state.
18 Faulkner, “Study of Composite Joint Strength with Carbon Nanotube Reinforcement,” 19.
27
C. CAL
CULATIONS
In order to calculate the Mode II critical strain energy release rate, GII, a
compliance method was used, which is based on the slope of the force versus
displacement load obtained during testing, i.e., a linear slope before crack propagation.
Once the compliance is obtained, the following equation is used to calculate GII:19
9a 2 Pc 2C
GII 
2b(2 L3  3a 3 )
The initial crack length (a), coupon width (b), and the span length (2L) are all
dependent on coupon geometry pre-determined prior to the start of the test. The critical
load, Pc, was determined based on the local maximum or slope change in the load versus
displacement curve, as well as observation.
Compliance was determined after the
completion of the test by taking the inverse of the slope of the load versus displacement
prior to crack propagation.
The compliance method is actually one of two ways to calculate GII. The first
method based on the Modified Beam Theory method requires material properties to be
known, as well as precise measurement of height and thickness of the samples. The
second method, the compliance approach, was chosen as it does not require material
properties be known.
Although it could easily be determined what these material
properties are, they vary depending on the CNT included and the thickness of coupon.
The compliance approach indirectly measures the material properties when calculating
the compliance.
No additional calculations were required for Phase III and IV resistance testing.
All data collect was already in the desired form of resistance measurements. For Phase V
the slope of the line formed by data points on the crack length versus resistance reading
graphs were calculated.
19M. Todo, T. Nakamura, and K. Takahashi, “Effects of Moisture Absorption on the Dynamic
Interlaminar Fracture Toughness of Carbon/Epoxy Composites,” Journal of Composite Materials 34(2000):
630–648.
28
V.
A.
RESULTS AND DISCUSSION
PHASE I: FAMILIRIZATION
During this phase, several samples were made although very few were useable.
The first four sets of samples constructed were four times thicker than those ultimately
tested in follow-on phases. The first two of these samples did not turn out properly due
to a leak in the vacuum bag seal that, even after repair, allowed too much air in to salvage
the sample. The third sample did not turn out to be good due to the thickness of the
sample, and the inability of the pump to completely pull the resin through the entire
sample. To fix this problem, on the fourth sample an extra strip of distribution media was
used in the middle bottom portion of the sample. This allowed for three different paths
for the resin to follow, ensuring the middle of the sample was thoroughly infused with
resin. On most days this sample would have turned out correctly, but it never gelled in
time allowing air to enter in. This is when it was discovered that, for most of the year in
Monterey, CA, N-dimthylaniline (DMA) is required to ensure proper resin cure times.
The last two samples constructed in Phase I were used to ensure that all
procedures consistently worked. With the use of DMA included in the resin and hardener
mixture, all samples were made successfully. These samples were not put through Mode
II testing, but were used to test new cutting techniques. Normally, samples of this nature
are cut into coupons using a water jet, but after trial and error it was determined that a
band saw with the correct blade can also cut composite samples.
B.
PHASE II: CO-CURED VS. TWO-STEP CURED
The first coupon tested was a two-step cured coupon with a 2.6 cm initial crack
length, span length of 15 cm, and width of 2.4 cm. The load was applied in the middle of
the span length at a location of 4.9 cm from the crack tip. Prior to signs of crack
propagation, this coupon failed at the point of the load application. This was not what
was desired, and so the speed was slowed down to 0.5 mm/min from 1 mm/min. This
was done in order to ensure that bending stress within the sample were not larger than
failure stresses.
29
The second coupon, tested with the new test speed, was also a two-step cured
coupon with the same geometry. Again, the coupon failed at the point of load application
prior to any crack propagation. Since the crack length was relatively small compared to
the span length, in both tests the load was applied too far away from the crack tip.
Bending stresses were reached prior to the onset of crack propagation.
To correct for this problem, the base supports on the Instron were moved closer
together to reduce the span length by one cm. The next coupon tested was two-step
cured, with the same crack length and width. Again the coupon failed at the point of load
application. A fourth two-step cured coupon was tested after moving the base supports in
an additional one cm. This time the crack began to propagate prior to buckling failure at
the point of load application. This failure corresponded to a span length of 13 cm, and
placed the load application much closer to the crack edge.
The span length of 13 cm was used for two more two-step cured coupons without
experiencing any more failures due to bending. However, this was not the case for all
coupons, as again bending failure was experienced on another coupon. The span length
of 13 cm was found to be the borderline length where either bending or crack propagation
could occur. A new span length of 12 cm was found to be an ideal length, and all other
coupons were tested using this span length.
The ratio of crack length to one-half the span length for the 15 cm span length
was found to be 0.35. The ratio for a span length of 13 cm, which appeared to be the
borderline span length, was 0.4. This shows that, for this particular sample set, a ratio of
crack length to one-half the span length for the test speed of 0.5 mm/min should be
greater than 0.4. For this particular sample set, such a ratio ensures that bending stresses
are not exceeded prior to the stress required for crack propagation.
Once a set span length was acquired, test results showed that there was a slight
increase in GII for two-step cured coupons over that of co-cured coupons. Figure 23
shows the normalized average values of GII for Phase II coupons, including the standard
deviation among the coupons tested. This data indicates that the two-step cured sample
sets had GII values 3.8% higher than the co-cured sample sets. The actual values of each
30
coupon can be seen in Appendix A. From this it is important to note that two-step cured
coupons had a wider range of GII values, but most of them were higher than that of the
co-cured coupons.
Figure 23.
Normalized Average Values of GII for Phase II
Upon further investigation, it was observed that the crack propagation was similar
for both the two-step cured and co-cured sample sets. For both cases, the crack initially
propagated from the built-in crack tip and ran along the centerline of the coupon
perpendicular to the load application.
Figures 24 and 25 show the path of crack
propagation as described.
31
Figure 24.
Figure 25.
Crack Propagation Path for a Co-Cured Coupon
Crack Propagation Path for a Two-Step Cured Coupon
After testing was complete, coupons in which crack propagation occurred were
pulled apart to inspect the cracked surface.
32
Both the co-cured and two-step cured
coupons experienced the same type of failure. In some areas, the joint interface bond was
broken through the resin, while in others the resin was pulled away from the fibers, as
shown in Figures 26 and 27.
Figure 26.
Figure 27.
Surface Crack Propagation Path for a Co-Cured Coupon
Surface Crack Propagation Path for a Two-Step Cured Coupon
Since both the co-cured and two-step cured samples failed in a similar manner, a
probable cause for the two-step cured higher GII values is related to the VARTM process.
33
When making two-step cured samples, the surface of the bottom resin layer is sanded and
cleaned carefully with acetone. During this process micro-scale defects, like voids in the
resin layer, are reduced, allowing for a stronger boundary interface to form between the
top and bottom fiber layers.
C.
PHASE III: CARBON COMPOSITE RESITANCE TESTING
This phase began with Mode II testing of all carbon composite coupons
containing CNTs. Based on Phase II results, a ratio of crack length to one-half the span
length of greater than 0.4 was desired; as a result, the initial crack length was chosen to
be 4 cm, with a span length of 16 cm, and width of 2.4 cm. These geometry parameters
resulted in a ratio of 0.5, which with a Mode II test speed of 1 mm/min, resulted in
coupon failure through crack propagation.
Prior to the start of testing each coupon was measured to determine its resistivity
for baseline comparisons. Each of these starting resistance readings can be seen in
Appendix B, and shows a varying degree of starting resistances. This is due to the
unevenly spread CNT, directly resulting from the dispersion technique used during the
VARTM process. Each value recorded however, was constant to within a tenth of an
ohm, and was read several different times before recording values.
During the actual testing, values of the resistance readings were recorded
manually at 30 second intervals. These values varied little from the initial readings
throughout the entire test. In fact most of the averages of these readings, with the
exception of those coupons with higher initial resistance readings, matched within 14%
of the initial resistance readings. Even when the sample cracked and continued to crack,
the resistance readings stayed constant varying only a few ohms at a time. The averages
resistance readings throughout the test are summarized in Appendix B.
When the test was complete the sample was left in the bent position shown in
Figure 28. The readings taken in the bent position were again constant, only fluctuating
to the tenth of an ohm, and within 4% of the initial resistance values. When the coupons
were released from this bent position, the resistance readings for all coupons increased,
and are listed in Appendix B. Again the variance in the increase percentage can be
34
contributed to the CNT dispersion method used during the VARTM process. On average
the increase in resistance readings for carbon composite coupons with a layer of CNTs
was 15.7%. This increase in resistance is what is desired in order to use CNTs as a
possible NDT method.
Figure 28.
Carbon Fiber Mode II Resistance Testing Bent Position
After experiencing such positive results from the carbon composite CNTs
reinforced coupons, the pure carbon composite coupons were tested. The first coupon
tested was setup with the same geometric parameters and Mode II test speed. However,
since the speed was faster than that used in Phase II, the coupon failed through bending in
the middle at the point of load application. Another pure carbon composite coupon was
tested to ensure that these test parameters were faulty for pure carbon composite coupons.
This second coupon failed in the same manner, and as a result the geometric parameters
were changed for the rest of the coupons. The remaining eight coupons were tested
having an initial crack length of 4 cm, a span length of 15 cm, and width of 2.4 cm.
Again prior to the start of testing, each coupon was measured to determine its
resistivity for baseline comparisons. Each of these starting resistance readings can be
seen in Appendix C, and shows a varying degree of starting resistances. For the pure
carbon composite coupons the resistance readings were very inaccurate and by no means
35
repeatable. Each time the coupons were hooked up to the multi-meter they started at a
given value and fluctuated widely.
After fluctuating for a little bit, all coupons’
resistance readings began to steadily increase, acting as a capacitor.
This was an
unexpected result, but validated the conductive behavior of CNTs when included in
carbon composites.
For pure composite coupons the resin, which is non-conductive in nature, is what
is being measured for resistance. Unfortunately, instead of acting as an open circuit, as
would be expected of a non-conductive material, the resin layer behaved as a capacitor.
Since the thickness of the layer of resin, compared to that of the surrounding carbon, was
thin, the carbon was able to sense some of the electricity being run through the stainless
steel. This flow of electricity was then transferred to the resin. The resin was charged by
the surrounding carbon, and in essence became a capacitor.
During Mode II testing of the pure carbon composite samples, resistance readings
were recorded manually at 30 second intervals. These values typically started high and
as the load was increased, they gradually decreased. For each coupon tested, at a certain
point during the Mode II testing, the values became steady and unchanging. These values
were extremely low in comparison to the initial fluctuating values experienced prior to
testing. The low steady resistance readings were a result of the sample being placed
under stress. When placed under stress, the carbon was not able to charge the resin layer
as it had before. Instead, the resin layer was compressed and too small for the carbon to
charge. The low readings, were in fact, those of the carbon layers.
When the test was complete, the coupon was left in the bent position and a
resistance reading was recorded. Readings taken in the bent position, for all pure carbon
composite coupons, were steady only fluctuating to the tenth of an ohm. These resistance
readings were extremely low compared to the initial readings taken, and are given in
Appendix C. Also given in this appendix are the average resistance values felt during
Mode II testing.
When the coupons were released, and returned to a flat position, an additional
resistance reading was taken. The resistance, for all pure carbon composite coupons,
36
increased, and then steadily began to climb, again taking on the behavior of a capacitor.
The values recorded in Appendix C are the values taken upon initially being returned to
the flat position. All readings are the baseline from which the resistance started to
quickly grow. Thus, unlike the carbon composite coupons with CNTs, the pure carbon
composite coupons respond poorly to the electrical resistance test. NDT could not be
used for pure carbon composite coupons, but is a valuable technique for carbon
composite reinforced with CNT.
To ensure that this was still a valuable use for the strengthening of composites at
areas of high concentration, and verify the results of previous research done at NPS, the
GII values for both the carbon composites with and without CNTs were calculated.
Figures 29 and 30 show each coupon’s load versus extension graphs used to calculate the
required GII values. The two graphs show that carbon composites hold load the same
way for both with and without CNTs, however, the crack location for composites with
CNTs is prolonged. Those coupons with CNTs also were able to reach higher loads
before complete load failure. This was verified by the test results that showed there was
an increase in GII for carbon composite coupons with CNTs over that of pure carbon
composite coupons. Figure 31 displays the normalized average values of GII for Phase III
coupons, along with the respective standard deviations. This data indicates that the
carbon composite sample sets reinforced with CNTs had GII values 20% higher than the
pure carbon composite sample sets. The actual values of each coupon can be seen in
Appendix D. From this it is important to note that each sample set had a similar standard
deviation, and that the highest value of the pure carbon sample set was barely higher than
the lowest sample set coupon reinforced with CNT.
37
Figure 29.
Figure 30.
Mode II Graph of Carbon Composites With CNT
Mode II Graph of Carbon Composites Without CNT
38
Figure 31.
Normalized Average Values of GII for Phase III
Based on previous research conducted at NPS20, the layer of CNTs included
within the carbon composite acted as expected. The pure carbon composite samples
experienced crack propagation at the initial crack tip, followed by propagation through
the joint interface. Carbon composites with CNTs, first experienced cracking at areas
away from the crack tip, i.e., weaker zones. These cracks then propagated back towards
the initial crack tip. These different crack propagations can be verified by observing the
surface of the joint interfaces where cracking occurred. Figure 32 shows the relatively
smooth joint interface of a pure carbon sample, with little fibers broken. This is a result
of the crack propagating through the joint interface. Figure 33 on the other hand shows
the rougher joint interface of the carbon composite containing CNTs. The rough surface
has CNTs on both sides, as well as several areas were the crack propagated back to the
initial tip through fibers. The crack was forced to propagate through the fibers due to the
CNTs being located in the joint interface strengthening it and making it resistance to
crack propagations.
20 Faulkner, “Study of Composite Joint Strength with Carbon Nanotube Reinforcement,” 27–30.
39
Figure 32.
Figure 33.
Surface Crack Propagation Path of Carbon Composite Without CNT
Surface Crack Propagation Path of Carbon Composite With CNT
40
D.
PHASE IV: FIBERGLASS COMPOSITE RESISTANCE TESTING
This phase began with Mode II testing of all fiberglass composite coupons
containing CNTs. Based on the results of Phase II, the initial crack length was chosen to
be 4 cm, with a span length of 16 cm, and width of 2.4 cm. These geometry parameters
along with a Mode II test speed of 1 mm/min, resulted in coupon failure through crack
propagation.
Prior to the start of testing each coupon was measured to determine its resistivity
for baseline comparisons.
Unfortunately, only four of the coupons made actually
registered any resistance on the multi-meter. An advantage to using fiberglass for testing
is that the CNTs inside the fiberglass composite could easily be seen. For the six
coupons that did not conduct, large areas within the coupons that were devoid of CNTs
could be detected as shown in Figure 34. Each of the four coupons that did conduct had a
visual path of CNTs that were continuous throughout the entire length of the coupon, as
displayed in Figure 35. This shows that in order for CNTs to be effective, they must be
touching. At the same time though, it also shows that if the CNTs are touching, they can
be effective even in non-conductive base composite materials. In order to ensure that
CNTs are touching, a better method for dispersion during the VARTM process should be
developed. A better method of dispersion would result in no open gaps, as experienced
in this particular sample set. All of the samples had large areas on one side or the other
were CNTs could not be seen. Figure 32 and 33 both shown areas on the top which are a
result of the Teflon used to create the cracks, however, the one in the middle on the
bottom are in fact devoid of CNTs. Again this is a result of the VARTM process and the
uneven dispersion method used.
41
Figure 34.
Fiberglass Coupon With Gaps in the Layer of CNTs
Figure 35.
Fiberglass Coupon With Continuous Layer of CNTs
Even though only four of the coupons were conductive, all coupons containing
CNTs were put through Mode II testing and values of the resistance readings were
recorded manually at 30 second intervals. The six coupons that initially did not conduct,
still registered no readings during the entire test. The values for the four conducting
fiberglass coupons, although much higher than those obtained for the carbon composite
coupons in Phase III, showed the same steady trend. During the test the resistance
readings varied little from the initial readings, and matched within 6%. Even when the
sample cracked and continued to crack, the resistance readings stayed constant varying
only a few ohms at a time, again consistent with Phase III carbon composite coupons
with CNTs. The average resistance readings throughout the test are summarized in
Appendix E.
42
When the test was complete the sample was left in the bent position as was done
during Phase III. The six coupons that were non-conductive still registered no resistance;
however the remaining four continued to give constant resistance readings. The readings
in the bent position were constant, but all readings had increased from the initial values,
some by as much as 30%. When the coupons were released from this bent position, the
four conducting fiberglass coupon’s resistance readings continued to increase while the
non-conduction fiberglass coupon’s remained unchanged. Both the bent and flat readings
for the four conducting fiberglass coupons are listed in Appendix E. Although each
coupon showed an increase in resistance, some showed higher percentages than others.
This variance can be contributed to the CNT dispersion method used during the VARTM
process. On average the increase in resistance readings for fiberglass coupons with CNTs
was 42.9%. Although much higher, this increase was consisted with Phase III results,
and is even more significant since it occurred in a non-conductive base material.
Next, the pure fiberglass coupons were tested. For these coupons resistance
readings were simple to take throughout the entire testing process. Each of the ten
coupons manufactured were tested and each one acted as an open circuit before, during
and after Mode II testing.
This was what was expected of a non-conducting base
composite material, and is exactly how the fiberglass composite samples with gaps in the
CNTs behaved. This is evidence that proofs further, that in order for CNTs to be
effective, they must be touching. When touching, CNTs will conduct and can be sensed
by a simple multi-meter.
Research, previously conducted with CNTs at the Naval Postgraduate School,
focused on the use of CNTs to strengthen carbon composite structures.21 Therefore it
was important to know if CNTs would strengthen fiberglass composites in the same way
that they do carbon composite structures. To do this, the GII values for both the fiberglass
composites with and without CNTs were calculated. Test results showed that there was
an increase in GII for fiberglass composite coupons with CNTs over that of pure
fiberglass coupons. Figure 36 displays the normalized average values of GII for Phase IV
21 Faulkner, “Study of Composite Joint Strength with Carbon Nanotube Reinforcement,”1–42.
43
coupons, along with the respective standard deviations. This data indicates that the
fiberglass composite sample sets with CNTs had GII values 35% higher than the pure
carbon composite sample sets.
The actual values of each coupon can be seen in
Appendix F. From this it is important to note that the highest value of the pure fiberglass
sample set was barely higher than the lowest sample set value for fiberglass with CNTs.
Mode II Average Values
1.2
Normalized GII
1.0
Fiberglass
With CNT
0.8
Fiberglass
Without
CNT
0.6
0.4
0.2
0.0
Figure 36.
Normalized Average Values of GII for Phase IV
When testing the carbon composites in Phase III, for both with and without CNTs,
the way in which they failed was expected based on previous research already
conducted.22 Fiberglass however, was surprising in its behavior both with and without
CNTs. During testing of fiberglass coupons with CNTs, a loud cracking sound could be
heard upon failure followed by a quick decrease in the loading. This can be seen in
Figure 37, which displays the load versus extension graph for all fiberglass coupons with
CNTs. The peak of each graph closely corresponds to the crack propagation point
observed visually, audibly, and graphically. This loud cracking sound was not observed
during testing of fiberglass composites without CNT, instead a soft crackling sound could
be heard. Also with the pure fiberglass coupons, after the crack could be visually and
22 Faulkner, “Study of Composite Joint Strength with Carbon Nanotube Reinforcement,” 22–31.
44
audibly verified, loads being applied still continued to climb. This can be shown in
Figure 38, which also displays the location where the crack could be seen and heard.
Load (kN)
Fiberglass With CNT
0.7
Coupon 1
0.6
Coupon 2
0.5
Coupon 4
Coupon 5
0.4
Coupon 6
0.3
Coupon 7
0.2
Coupon 8
0.1
Coupon 9
0
0
2
4
6
8
10
12
14
Coupon 10
Extension (mm)
Figure 37.
Figure 38.
Mode II Graph of Fiberglass Composites With CNT
Mode II Graph of Fiberglass Composites Without CNT
Differences in both the sound of failure, and crack propagation can be directly
contributed to the CNTs. In the non-reinforced samples, crack propagation began at the
tip of the initial crack, and continued to propagate through the joint interface, as shown in
45
Figures 39 and 40. This crack occurred early in the loading cycle and slowly propagated
while still maintaining an increasing load. For the fiberglass composites reinforced with
CNTS they too initially propagated from the crack tip through the joint interface,
however, at a certain point the crack took the path of least resistance under the layer of
CNTs, as shown in Figures 41 and 42. This result was widely observed in the CNT
reinforced samples, and is the reason for the loud crack sound heard.
Figure 39.
Figure 40.
Fiberglass Composites Without CNT Path of Crack Propagation Drawing
Fiberglass Composites Without CNT Path of Crack Propagation Picture
46
Figure 41.
Figure 42.
Fiberglass Composites With CNT Path of Crack Propagation
Fiberglass Composites With CNT Path of Crack Propagation Picture
After testing was complete for all phases of research, coupons in which crack
propagation occurred were pulled apart to inspect the cracked joint interface surface, and
verify crack propagation paths. When the fiberglass coupon with CNTs was pulled apart,
one side contained more CNTs than the other. Looking closer it could be seen that
initially the crack did propagate through the layer of CNTs, but then quickly took the path
of least resistance under the layer of CNTs through the fiberglass. The fiberglass coupon
without CNTs showed a slightly different crack propagation path. The joint interface
47
bond was broken through the resin by the crack propagation resulting in the resin being
pulled away from the fibers. Both surface interfaces are shown below in Figures 43 and
44.
Figure 43.
Figure 44.
Surface Crack Propagation Path of Fiberglass Composite With CNT
Surface Crack Propagation Path of Fiberglass Composite Without CNT
The variances in surfaces can also explain the differences in both the physical
observations, as well as the differences in the loads each sample set was able to carry.
48
The pure fiberglass composite acted as the two-step cured samples tested during Phase II.
The crack propagated through the joint interface, an area which was inherently stronger
due the VARTM process. This allowed for higher loads to be carried and slower crack
growth. The fiberglass with CNT acted more like the co-cured samples from Phase II.
Once the crack propagated into layers above or below that of the CNTs, it was
propagating through a weaker resin bond allowing faster crack propagation and lower
loads to be carried. This is ultimately why, although crack propagation was prolonged in
the fiberglass with CNTs, those without were still able to carry higher loads.
E.
PHASE V: RESIS TANCE RE LIABILITY AND CRACK GROWTH
RELATIONSHIP TESTING
This phase began with testing of the four fiberglass composite coupons containing
CNTs, from Phase IV, that resistance readings were able to be obtained. All coupons
tested were placed on the Instron with the same test setup from Phase III and IV. In other
words a span length of 16 cm, and width of 2.4 cm were still used. Before placing the
coupons into the machine however, the length of the crack resulting from Phase IV Mode
II tests were measured and recorded. Once loaded into the Instron, a load of 100 kN was
applied to the coupons so that the crack was stationary without growth and the
corresponding resistance readings were taken for both bent and unbent readings. This
was done at least three times for each sample. The resulting resistance readings can be
seen in Appendix G.
Although the readings vary from the cracked resistance readings taken in Phase
IV, shown in Appendix E, each coupon is consistent within itself, only varying by at most
6.35%. Again the difference between the different coupons can be attributed to the
uneven distribution of CNTs within the coupons. The readings also vary from that taken
in Phase IV due to the different placement of where the multi-meter is attached on the
sample. If this practice is to be used, the exact location of the test equipment placement
must be marked in order to ensure consistent readings from one test to the next.
After taking the consistency readings, the fiberglass coupons were then manually
loaded for crack growth using the Instron machine. Unfortunately, no useful information
49
was gathered from this step. Upon crack propagation, resistance readings jumped to over
1 MΩ. These high readings were indications that the CNTs were no longer touching and
the sample was now acting as an open circuit. In essence, the crack had severed the
continuous layer of CNTs and began to propagate below the layer of CNTs. This is what
was observed and discussed in the Phase IV results when the samples were pulled apart
for inspection.
The same two tests were then conducted using all carbon composite coupons
containing CNTs, from Phase III. The same geometry setup was used, and the length of
the crack resulting from Phase III Mode II tests was measured and recorded. This time a
load of 50 kN was applied to the coupons, as 100 kN may have been a cause of the quick
crack propagation experienced with the fiberglass coupon number four.
The
corresponding resistance readings for both the bent and unbent positions were taken.
This was done at least three times for each sample. The resulting resistance readings can
be seen in Appendix H.
As with the fiberglass composites, the readings for the carbon composites were
consistent with each other. The average change in resistance was 1.26% with the highest
resistance change being 8.77%. Any difference between the coupons can be attributed to
the uneven distribution of CNTs within the coupons. As was seen with the fiberglass, the
carbon readings also varied from those taken in Phase III. As already discussed, this is
due to the different placement of where the multi-meter is attached on the sample.
After taking the consistency readings, the carbon coupons were then manually
cracked using the Instron machine. Once the crack propagated, which was determined by
both sight and sound, the new crack length was measured, and the corresponding
resistance reading was taken. This was done repeatedly until the crack tip had reached
the point of load application, and it was no longer possible to further crack the coupons
with the Instron machine.
The resulting data was then plotted to determine any
relationship between change of crack length and change in resistance. Figure 45 shows
all the data collected for coupons with CNT on one large graph.
50
Figure 45.
Carbon Composite Resistance vs. Crack Length Graph For All CNT
Coupons
The above graph shows all the data minus the initial crack information. This was
done to be more consistent in calculations. The resistance readings for the initial crack
length of 4 cm were taken using a different test equipment position for the multi-meter,
and therefore were not valid for these calculations. Unfortunately, even with these
readings removed from the graph the data was still verily spread apart. This again is due
to the uneven dispersion of CNTs in each of the coupons.
In an attempt to find some relationship between changes of crack length to
changes of resistance, each coupon’s data was plotted on its own graph. Most of the data
followed a linear behavior, and so a linear regression was performed for each plot as
shown in Figures 46, 47, 48, 49, 50, 51, 52, 53, and 54. These figures show that no
standard slope could be found but an average was taken to be 13.68 Ohms/mm with a
standard deviation of 14.52 Ohms/mm.
51
Figure 46.
Carbon Composite Coupon 1 Resistance vs. Crack Length Graph
Figure 47.
Carbon Composite Coupon 2 Resistance vs. Crack Length Graph
52
Figure 48.
Carbon Composite Coupon 3 Resistance vs. Crack Length Graph
Figure 49.
Carbon Composite Coupon 4 Resistance vs. Crack Length Graph
53
Figure 50.
Carbon Composite Coupon 5 Resistance vs. Crack Length Graph
Figure 51.
Carbon Composite Coupon 6 Resistance vs. Crack Length Graph
54
Figure 52.
Carbon Composite Coupon 8 Resistance vs. Crack Length Graph
Figure 53.
Carbon Composite Coupon 9 Resistance vs. Crack Length Graph
55
Figure 54.
Carbon Composite Coupon 10 Resistance vs. Crack Length Graph
These figures all have the same general trends no matter what the starting crack
length and initial resistance reading. With each incremental increase in crack length, the
resistance values increased. Although it was difficult to predict how much the crack
would propagate each time it was loaded, the resistance never failed to increase, even
with the smallest increase in crack length. This increase in resistance is related to the fact
that the crack for a carbon composite with CNTs propagates through the layer of CNTs.
Thus, as the crack continues to propagate, the CNTs are separated from each other, and
their ability to conduct throughout the sample is decreased. The more holes in the layer
of CNTs, the harder it is to conduct, and thus an increase in resistance.
Although there seems to be a linear relationship, more testing needs to be done to
verify these findings. More data points need to be taken in order to truly determine if a
linear relationship is the correct one to assign to the resistance behavior of CNTs in
composite materials. For future work this data could be improved by ensuring even
dispersion of CNTs, designated test equipment positions for multi-meter, and a more
scientific method to predict crack propagation in intermediate steps.
56
VI. CONCLUSIONS
AND RECOMMENDATIONS
In conclusion, interface strength of woven fabric composite layers was studied
using Mode II fracture mechanics testing. Both carbon fiber and glass fiber composites
were used with the vinyl ester resin. Five phase of research were conducted, each
looking at different aspects of the interface strength of composite layers. First, the cocured composite interface strength was compared to that of the two-step cured interface
as used in the scarf joint technique. The test results showed that the two-step cured
interface was as strong as the co-cured interface, and the former had even higher fracture
toughness than the latter. The conclusion is that the two-step cured interface is slightly
better than the co-cured in terms of fracture toughness, however in terms of labor
intensiveness, co-cure would be simply preferable.
The second study applied carbon nanotubes to the composite interface using the
two-step cured technique. Mode II fracture testing was performed for the interface
containing CNTs. The results indicated a great improvement of the interface fracture
toughness due to CNTs for both carbon and fiberglass composites.
Finally, a study was conducted to detect interface crack growth using the CNTs
introduced at the interface. Because CNTs have high electric conductivity, the electric
resistance was measured through the interface. For fiberglass composites, due to their
unusual paths of crack propagation, only the initial failure was detected through
resistance. Carbon composites however, as the interface crack grew under loading, a
gradual increase of electric resistance was observed upon unloading. As a result, the
change of electric resistance in terms of crack length change was studied for carbon
composite materials. Unfortunately, due to uneven dispersion techniques, and other
testing procedures, it could only be determined that a linear relationship exists for these
carbon composite materials. The study did show that using CNTs in carbon composite
materials at a critical composite interface would not only strengthen its fracture
toughness, but also detect crack growth.
57
Further research is necessary to verify the above findings and conclusions. Tests
already conducted should be run again at different levels of CNTs concentrations, as well
as with a better CNTs dispersion method. This will lead to closer resistance readings
from coupon to coupon, and a more accurate crack length change to resistance
relationship. When conducting any resistance tests, exact locations for test equipment,
mainly the multi-meter clips, should be marked and used for all tests.
Furthermore, Mode I tests and Mixed Mode I-Mode II tests should be conducted
while measuring the conductivity of composite materials. In actual structures, the stress
will rarely be purely Mode II, and so all possibilities must be fully studied. Further
research is also needed to determine feasible manufacturing practices for local CNT
dispersion.
58
APPENDIX A: TWO-STEP CURED AND CO-CURED CRITICAL
STRAIN ENERGY RELEASE RATES (GII)
Two-Step Cured
Sample
2D
2E
2G
2H
2I
GIIC (N/m)
1.016E+03
9.533E+02
6.521E+02
7.168E+02
6.745E+02
C (m/N)
9.4697E-06
8.6957E-06
6.7340E-06
7.0771E-06
9.2507E-06
Pc (N)
713.901
721.556
608.57
643.331
545.831
L (cm)
6.5
6.5
6.0
6.0
6.0
a (cm)
2.6
2.6
2.6
2.5
2.5
b (cm)
2.40
2.40
2.40
2.40
2.40
Co-Cured
Sample
1C
1D
1E
1F
1G
1H
GIIC (N/m)
8.905E+02
8.741E+02
8.850E+02
5.933E+02
6.372E+02
7.500E+02
C (m/N)
1.0905E-05
1.1186E-05
1.1236E-05
9.0909E-06
9.2851E-06
9.7182E-06
59
Pc (N)
584.557
589.716
592.069
534.883
512.308
511.263
L (cm)
6.5
6.5
6.5
6.0
6.0
6.0
a (cm)
2.8
2.7
2.7
2.4
2.6
2.8
b (cm)
2.40
2.40
2.40
2.40
2.40
2.40
THIS PAGE INTENTIONALLY LEFT BLANK
60
APPENDIX B: CARBON COMPOSITE WITH CNT RESISTANCE
DATA PHASE III
Average
Sample
Number
Initial
Resistance
Resistance
During
(Ohms)
Testing
(Ohms)
Resistance
Resistance
in Bent
in Flat
Position
Position
(Ohms)
(Ohms)
Percent
Increase in
Resistance
1
173.3
173.1
173.5
182.2
5.14%
2
26.5
26.5
26.9
28.1
6.04%
3
49.3
49.2
49.2
51.2
3.85%
4
71.6
71.6
71.1
73.1
2.09%
5
232.5
234.9
235.2
241.4
3.83%
6
287.2
286.4
277.6
293.1
2.05%
7
74.5
85.2
75.2
123.4
65.64%
8
1081.0
1043.5
1046.0
1112.0
2.87%
9
455.6
281.1
148.5
622.8
36.70%
10
252.5
300.1
288.8
326.2
29.19%
270.4
255.2
239.2
305.4
15.7%
Group
Averages
61
THIS PAGE INTENTIONALLY LEFT BLANK
62
APPENDIX C: PURE CARBON COMPOSITE RESISTANCE DATA
PHASE III
Average
Sample
Number
Initial
Resistance
Resistance
During
(Ohms)
Testing
(Ohms)
Resistance
Resistance
in Bent
in Flat
Position
Position
(Ohms)
(Ohms)
Percent
Increase in
Resistance
1
91.25
25.93
16.61
76.60
-16.05%
2
9.16
4.27
3.67
6.23
-31.99%
3
1750.00
322.15
8.75
5500.00
214.29%
4
9.73
7.50
7.94
9.82
0.92%
5
18.10
10.64
5.05
18.10
0.00%
6
35.30
9.71
10.80
38.40
8.78%
7
453.00
8.82
5.21
71.30
-84.26%
8
435.20
58.56
4.30
454.00
4.32%
9
230.00
6.37
4.55
59.50
-74.13%
10
17900.00
23.00
7.25
1270.00
-92.91%
2093.17
47.69
7.41
750.40
-7.10%
Group
Averages
63
THIS PAGE INTENTIONALLY LEFT BLANK
64
APPENDIX D: CARBON COMPOSITE WITH AND WITHOUT CNT
CRITICAL STRAIN ENERGY RELEASE RATES (GII)
With CNT
Sample
1
2
3
4
5
6
7
8
9
10
GIIC (N/m)
1.069E+03
1.161E+03
1.056E+03
1.103E+03
1.208E+03
1.023E+03
1.272E+03
9.998E+02
1.244E+03
1.116E+03
C (m/N)
1.8797E-05
1.7953E-05
1.6129E-05
1.5898E-05
1.9084E-05
1.5361E-05
1.7123E-05
1.6835E-05
1.7483E-05
1.6367E-05
Pc (N)
480.229
512.076
515.066
530.225
506.515
519.672
548.763
490.636
537.113
525.651
L (cm)
8.0
8.0
8.0
8.0
8.0
8.0
8.0
8.0
8.0
8.0
a (cm)
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
b (cm)
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
GIIC (N/m)
8.392E+02
8.207E+02
9.486E+02
1.045E+03
7.502E+02
9.106E+02
1.042E+03
8.431E+02
C (m/N)
1.8519E-05
1.6892E-05
1.6313E-05
1.6892E-05
1.8692E-05
1.9920E-05
1.7483E-05
1.5504E-05
Pc (N)
395.551
409.557
448.06
462.116
372.25
397.264
453.597
433.296
L (cm)
7.5
7.5
7.5
7.5
7.5
7.5
7.5
7.5
a (cm)
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
b (cm)
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
Without CNT
Sample
3
4
5
6
7
8
9
10
65
THIS PAGE INTENTIONALLY LEFT BLANK
66
APPENDIX E: FIBERGLASS COMPOSITE WITH CNT
RESISTANCE DATA PHASE IV
Average
Sample
Number
Initial
Resistance
Resistance
During
(Ohms)
Testing
(Ohms)
Resistance
Resistance
in Bent
in Flat
Position
Position
(Ohms)
(Ohms)
Percent
Increase in
Resistance
1
38,120
38,572
39,950
44,550
16.87%
2
357,100
336,850
404,100
455,300
27.50%
4
73,090
74,531
88,100
146,500
100.44%
7
717,600
742,434
939,200
909,200
26.70%
296,477
298,096
367,837
388,887
42.9%
Group
Averages
67
THIS PAGE INTENTIONALLY LEFT BLANK
68
APPENDIX F: FIBERGLASS COMPOSITE WITH AND WITHOUT
CNT CRITICAL STRAIN ENERGY RELEASE RATES (GII)
With CNT
Sample
1
2
4
5
6
7
8
9
10
GIIC (N/m)
9.803E+02
1.109E+03
1.054E+03
7.802E+02
9.257E+02
1.099E+03
1.084E+03
8.641E+02
8.978E+02
C (m/N)
1.3106E-05
1.2516E-05
1.2063E-05
1.3986E-05
1.1148E-05
1.2788E-05
1.1481E-05
1.2610E-05
1.3986E-05
Pc (N)
550.607
599.307
595.245
475.502
580.161
590.156
618.6
527.009
510.099
L (cm)
8.0
8.0
8.0
8.0
8.0
8.0
8.0
8.0
8.0
a (cm)
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
b (cm)
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
GIIC (N/m)
6.181E+02
6.142E+02
7.960E+02
5.929E+02
6.796E+02
6.245E+02
4.611E+02
6.594E+02
6.475E+02
6.594E+02
C (m/N)
1.2315E-05
1.2121E-05
1.2392E-05
1.0091E-05
1.0395E-05
1.0604E-05
9.4162E-06
9.4877E-06
8.6505E-06
9.2081E-06
Pc (N)
451.026
453.201
510.283
450.409
475.106
450.901
378.589
451.02
430.048
420.631
L (cm)
8.0
8.0
8.0
7.5
7.5
7.5
7.0
7.0
6.5
6.5
a (cm)
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
b (cm)
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
2.40
Without CNT
Sample
1
2
3
4
5
6
7
8
9
10
69
THIS PAGE INTENTIONALLY LEFT BLANK
70
APPENDIX G: FIBERGLASS COMPOSITE WITH CNT
RESISTANCE DATA PHASE V
Each coupon was tested at least three times using the following procedure.
1) Measure the crack length and initial resistance reading
2) Load and unload the coupon allowing no crack to propagate
3) Measure the resulting resistance reading
Each of the rows below represents the different trial runs for each sample.
Coupon 1:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 6.4 44300 100
44500
0.45% 6.4 44500 100
44600
0.22% 6.4 44600 100
44700
0.22% Average: 0.30%
Coupon 2:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 6.2 771100 100
820100
6.35% 6.2 739000 100
739200
0.03% 6.2 739200 100
709000
4.09% Average: 3.49%
Coupon 4:
New Crack Initial Resistance Load Final Resistance Length (cm) Reading (ohms) (N) Reading (ohms) 6.5 285500 100 over 1 MΩ 71
Coupon 7:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 6 912600 100
891200
2.34% 6 914400 100
908200
0.68% 6 904600 100
871100
3.70% Average:
2.24%
72
APPENDIX H: CARBON COMPOSITE WITH CNT RESISTANCE
DATA PHASE V
Each coupon was tested at least three times using the following procedure.
1) Measure the crack length and initial resistance reading
2) Load and unload the coupon allowing no crack to propagate
3) Measure the resulting resistance reading
Each of the rows below represents the different trial runs for each sample.
Coupon 1:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 5.7 182.1
50
182.3
0.11% 5.7 180.8
50
181.8
0.55% 5.7 181.5
50
181.9
0.22% Average:
0.29%
Coupon 2:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 5.5 28.9
50
28.8
0.35% 5.5 28.8
50
28.7
0.35% 5.5 28.8
50
28.9
0.35% Average:
0.35%
Coupon 3:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 6.3 51.3
50
51.7
0.78% 6.3 50.5
50
51.6
2.18% 6.3 51.2
50
51.6
0.78% 6.3 51.2
50
51.5
0.59% Average:
1.08%
73
Coupon 4:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 5.8 67.2
50
67
0.30% 5.8 67
50
66.9
0.15% 5.8 67
50
67.1
0.15% Average:
Coupon
0.20%
5:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 5.6 365.1
50
370.1
1.37% 5.6 372.2
50
371.1
0.30% 5.6 370.1
50
371.4
0.35% Average:
Coupon
0.67%
6:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 6.1 93.5
50
101.7
8.77% 6.1 108.4
50
108.6
0.18% 6.1 99.4
50
103.2
3.82% Average:
4.26%
Coupon 8:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 7.2 839.2
50
836.7
0.30% 7.2 840.1
50
846
0.70% 7.2 847.7
50
847.7
0.00% Average:
Coupon
0.33%
9:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 6.6 157.7
50
163
3.36% 6.6 161.2
50
159.1
1.30% 6.6 160.5
50
160.5
0.00% Average:
1.55%
74
Coupon 10:
New Crack Initial Resistance Load Final Resistance Percentage Length (cm) Reading (ohms) (N) Reading (ohms) Change 6.6 276.5
50
278.1
0.58% 6.6 276.1
50
276.6
0.18% 6.6 276.6
50
256.6
7.23% Average: 2.66%
75
THIS PAGE INTENTIONALLY LEFT BLANK
76
LIST OF REFERENCES
Callister, William D., Jr. Materials Science and Engineering: An Introduction, 7th ed.,
New York: John Wiley and Sons, Inc, 2007.
Chou, Tsu-Wei and Erik T. Thosetenson. Carbon Nanotube/Vinyl Ester Nanocomposites
for in Situ Sensing. September 17–19, 2008. University of Maryland University
College, Adelphia, MD. Office of Naval Research Solid Mechanics Program
Review Meeting: Marine Composites and Sandwich Structures.
Dharap, P., Z. Li, S. Nagarajaiah, and E.V. Barrera. 2004. Nanotube film based on singlewall carbon nanotubes for strain sensing. Nanotechnology 15.
Faulkner, Susan. “Study of Composite Joint Strength with Carbon Nanotube
Reinforcement,” Master’s thesis, Naval Postgraduate School, 2008.
Harris, P.J.F. 2004. Carbon Nanotube Composites. International Materials Review 49.
Kang, I., M. J. Schulz, J. H. Kim, V. Shanov, and Shi, D. 2006. A carbon nanotube strain
sensor for structural health monitoring. Smart Materials and Structures 15.
Kwon, Y. W., R. Slaff, S. Bartlett, and T. Greene. 2008. Enhancement of Composite
Scarf Joint Interface Strength through Carbon Nanotube Reinforcement. Journal
of Materials Science.
Live Journal. Definition of a Nanotube. March 12, 2009.
http://fullerenes.livejournal.com/ (accessed September, 9 2009).
Mouritz, A.P., E. Gellert, P. Burchill, and K. Challis. July 2001. Review of Advanced
Composite Structures for Naval Ships and Submarines. Composite Structures 53.
Nofar, M., S.V. Hoa, and M.D. Pugh. 2009. Failure Detection and Monitoring in Polymer
Matrix Composites Subjected to Static and Dynamic Loads Using Carbon
Nanotube Networks. Composites Science and Technology.
Saito, R. and M. S. Dresselhaus. Physical Properties of Carbon Nanotubes. 1998.
Imperial College Press.
The Venton Research Group. Development of Carbon Nanotube Modified
Microelectrodes. n.d. http://www.faculty.virginia.edu/ventongroup/nanotube.html
(accessed September 9, 2009).
Thostenson, E.T., and T.W. Chou. 2006. Carbon Nanotube Networks: Sensing of
Distributed Strain and Damage for Life Prediction and Self Healing. Advanced
Materials 18.
Todo, M., T. Nakamura, and K. Takahashi. 2000. Effects of Moisture Absorption on the
Dynamic Interlaminar Fracture Toughness of Carbon/Epoxy Composites. Journal
of Composite Materials 34.
77
Weber, I., and P. Schwartz. 2001. Monitoring Bending Fatigue In Carbon-Fibre/Epoxy
Composite Strands: A Comparison Between Mechanical and Resistance
Techniques. Composites Science and Technology 61.
Wong, M., M. Paramsothy, X.J. Xu, Y. Ren, S. Li, and K. Liao. December 2003.
Physical Interactions at Carbon Nanotube-Polymer Interface. Polymer 44.
Zhang, W., V. Sakalkar, and N. Koratkar. 2007. In Situ Health Monitoring and Repair In
Composites Using Carbon Nanotube Additives. Applied Physiscs Letters 91.
78
INITIAL DISTRIBUTION LIST
1.
Defense Technical Information Center
Ft. Belvoir, Virginia
2.
Dudley Knox Library
Naval Postgraduate School
Monterey, California
3.
Graduate School of Engineering and Applied Sciences
Naval Postgraduate School
Monterey, California
4.
Joe Johnson
Integrated Composites Inc.
Marina, California
5.
John Dickie
Integrated Composites Inc.
Marina, California
6.
John McWaid
Integrated Composites Inc.
Marina, California
7.
Ray Uncangco
Integrated Composites Inc.
Marina, California
8.
Professor Young W. Kwon
Naval Postgraduate School
Monterey, California
9.
Major Randall D Pollak
Naval Postgraduate School
Monterey, California
10.
Professor and Chairman Knox T. Millsaps
Naval Postgraduate School
Monterey, California
11.
Lieutenant Mollie A. Bily
Naval Postgraduate School
Monterey, California
79
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement